Windows live audio translator

Translate computer audio live on Windows.

The Windows companion is built for selected playback audio: webinars, browser audio, meeting sound, and translated output to the device you choose.

Download Chuchotage Supported platforms

Use caseA webinar, browser video, Teams-style call, or training audio playing on a Windows PC.

ArchitectureElectron UI with a .NET bridge for WASAPI loopback capture, Realtime Translation, and selected playback output.

StatusImplemented as a Windows companion in the repo; public Windows download is not posted yet.

What it does

The Windows companion captures audio from a selected Windows playback device, converts it to mono PCM16, streams it to OpenAI Realtime Translation, and plays translated audio to a selected output device.

It is intended to run as a normal user app without admin elevation, bundled virtual audio drivers, a Teams integration, or a Chuchotage backend for normal translation.

Screenshots

Current app screens, not mockups.

These current app screens show the shared Chuchotage flow while the Windows companion uses a desktop-specific capture backend.

Chuchotage listening screen showing automatic source language detection and stop translation control. — Automatic source-language detection while listening.

Chuchotage settings screen for choosing the output language. — Choose the language you want to hear.

Chuchotage credential screen for API key and ChatGPT sign-in options. — Credential setup stays device-local.

Platforms

Supported platforms today.

Windows 10/11

Companion work targets Windows 10 2004 or newer and Windows 11. Same-device exclusion behavior is best on Windows 10 Build 20348 or newer.

macOS

Mac app is the closest public desktop path today, with playback-audio capture on supported macOS versions.

iPhone and iPad

Public App Store app for microphone listen-along, not Windows computer-audio capture.

Android

Native Android app has a Device audio path, but Google Play release is not public yet.

Audio sources

Headphones, mics, and app audio.

Selected playback device

Capture the output where your browser, webinar, meeting app, or media is playing.

Translated playback device

Choose a separate output device when possible, usually headphones, to avoid translating Chuchotage's own output.

Single headset

The app can try process-loopback exclusion on supported Windows builds, but separate devices are cleaner when available.

Virtual drivers

Chuchotage does not install or require virtual audio drivers. Existing admin-approved virtual devices can be used manually.

Limits

What does not work yet.

Public Windows installer/download is not posted yet.

Process-loopback exclusion requires Windows 10 Build 20348 or newer.

Bluetooth headset routes and meeting apps can behave differently across machines.

No Teams bot, meeting recorder, cloud transcript archive, or admin-only driver install.

Privacy shape

Personal translation without an ad or analytics layer.

No ads and no analytics SDKs in the app.

Credentials and app preferences are stored on the device using the platform secure-storage path.

During normal API-key or ChatGPT use, selected audio and translation settings are sent from the app to OpenAI while translation is active.

Chuchotage does not run a hosted audio relay and does not keep transcript history.

Pick the route that matches what you need to translate.

Live speech translation Meeting audio translation Headphone translation Android device audio Private speech translation AirPods alternative Google Translate alternative

Download

Get Chuchotage for the platforms available today.

Download